Can dimedone be used to study selenoproteins? An investigation into the reactivity of dimedone toward oxidized forms of selenocysteine

Article abstract of DOI:10.1002/pro.3390

Dimedone is a widely used reagent to assess the redox state of cysteine-containing proteins as it will alkylate sulfenic acid residues, but not sulfinic acid residues. While it has been reported that dimedone can label selenenic acid residues in selenoproteins, we investigated the stability, and reversibility of this label in a model peptide system. We also wondered whether dimedone could be used to detect seleninic acid residues. We used benzenesulfinic acid, benzeneseleninic acid, and model selenocysteine-containing peptides to investigate possible reactions with dimedone. These peptides were incubated with H₂O₂ in the presence of dimedone and then the reactions were followed by liquid chromatography/electrospray ionization mass spectrometry (LC/ESI-MS). The native peptide, H-PTVTGCUG-OH (corresponding to the native amino acid sequence of the C-terminus of mammalian thioredoxin reductase), could not be alkylated by dimedone, but could be carboxymethylated with iodoacetic acid. However the “mutant peptide,” H-PTVTGAUG-OH, could be labeled with dimedone at low concentrations of H₂O₂, but the reaction was reversible by addition of thiol. Due to the reversible nature of this alkylation, we conclude that dimedone is not a good reagent for detecting selenenic acids in selenoproteins. At high concentrations of H₂O₂, selenium was eliminated from the peptide and a dimeric form of dimedone could be detected using LCMS and ¹H NMR. The dimeric dimedone product forms as a result of a seleno-Pummerer reaction with Sec-seleninic acid. Overall our results show that the reaction of dimedone with oxidized cysteine residues is quite different from the same reaction with oxidized selenocysteine residues.

Full text of DOI:10.1002/pro.3390

Article
Protein Science
DOI 10.1002/pro.3390
Can Dimedone be Used to Study Selenoproteins? An
Investigation into the Reactivity of Dimedone Towards
Oxidized Forms of Selenocysteine†
N. Connor Payne‡, a, b, Drew R. Barber‡, a, c, Erik L. Ruggles‡, and Robert J.
Hondal*‡
aThese authors made equal contributions to this manuscript.
bNCP was supported in part by the University of Vermont Office of Undergraduate Research in
the form of a Summer Research Award.
cDRB was supported by National Institutes of Health Training Grant T32 HL07594 administered
by Dr. Kenneth G. Mann and Dr. Robert J. Kelm.
†These studies were supported by National Institutes of Health Grant GM094172 to RJH.
‡Department of Biochemistry, 89 Beaumont Ave, Given Building Room B413, Burlington, VT
05405
*To whom correspondence should be addressed: Department of Biochemistry, University of
Vermont, College of Medicine. 89 Beaumont Ave, Given Building Room B413, Burlington, VT
05405. Tel: 802ꢀ656ꢀ8282. FAX: 802ꢀ656ꢀ8220. Eꢀmail: Robert.Hondal@uvm.edu
Running Title: The use of dimedone to study selenoproteins
Key words: sulfenic acid, sulfinic acid, selenenic acid, seleninic acid, dimedone, Pummerer
reaction, selenocysteine
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/pro.3390
© 2018 The Protein Society
Received: Dec 28, 2017; Revised: Feb 12, 2018; Accepted: Feb 13, 2018
This article is protected by copyright. All rights reserved.

Protein Science
Page 2 of 63
2
ABBREVIATIONS AND TEXTUAL FOOTNOTES
dSec, selenocysteine; H2O2, hydrogen peroxide; Cys, cysteine; GSH, glutathione; dimedone, 5,5ꢀ
dimethylcyclohexaneꢀ1,3ꢀdione; MS, mass spectrometry; LCMS, liquid chromatographyꢀmass
spectrometry; Ser, serine; mTrxR, mammalian thioredoxin reductase; Trx, thioredoxin; Secꢀ
ꢀ
SeO2 , selenocysteineꢀseleninic acid; Ala, alanine; βME, βꢀmercaptoethanol; TCEP, tris(2ꢀ
carboxyethyl)phosphine; HED, hydroxyethyl disulfide; TLC, thinꢀlayer chromatography; NMR,
nuclear magnetic resonance; HPLC, highꢀperformance liquid chromatography; EDTA,
ethylenediaminetetraacetic acid; NADPH, nicotinamide adenine dinucleotide phosphateꢀreduced;
TFA, trifluoroacetic acid; TIS, triisopropylsilane; Et2O, diethyl ether; APCI, atmosphericꢀ
pressure chemical ionization; ESI, electrospray ionization; DTNP, 2,2′ꢀdithiobis(5ꢀ
nitropyridine); 5ꢀNpys, 5ꢀnitroꢀ2ꢀpyridinesulfenyl protecting group; PhSeO2H, benzeneseleninic
acid; MeSeO2H, methaneseleninic acid; Gly, glycine; DTT, dithiothreitol; Mob, pꢀ
methoxybenzyl protecting group.
eNote: The wild type mammalian TrxR is abbreviated as mTrxRꢀGCUG to indicate the amino
acid composition of the last four amino acids. Likewise, mutants of mTrxR are abbreviated as
mTrxRꢀAA1AA2AA3AA4.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 3 of 63
Protein Science
3
ABSTRACT
Dimedone is a widely used reagent to assess the redox state of cysteineꢀcontaining proteins as it
will alkylate sulfenic acid residues, but not sulfinic acid residues. While it has been reported that
dimedone can label selenenic acid residues in selenoproteins, we investigated the stability, and
reversibility of this label in a model peptide system. We also wondered whether dimedone could
be used to detect seleninic acid residues. We used benzenesulfinic acid, benzeneseleninic acid,
and model selenocysteineꢀcontaining peptides to investigate possible reactions with dimedone.
These peptides were incubated with H2O2 in the presence of dimedone and then the reactions
were followed by liquid chromatography/electrospray ionization mass spectrometry (LC/ESIꢀ
MS). The native peptide, HꢀPTVTGCUGꢀOH (corresponding to the native amino acid sequence
of the Cꢀterminus of mammalian thioredoxin reductase), could not be alkylated by dimedone, but
could be carboxymethylated with iodoacetic acid. However the “mutant peptide”, Hꢀ
PTVTGAUGꢀOH, could be labeled with dimedone at low concentrations of H2O2, but the
reaction was reversible by addition of thiol. Due to the reversible nature of this alkylation, we
conclude that dimedone is not a good reagent for detecting selenenic acids in selenoproteins. At
high concentrations of H2O2, selenium was eliminated from the peptide and a dimeric form of
dimedone could be detected using LCMS and 1H NMR. The dimeric dimedone product forms as
a result of a selenoꢀPummerer reaction with Secꢀseleninic acid. Overall our results show that the
reaction of dimedone with oxidized cysteine residues is quite different from the same reaction
with oxidized selenocysteine residues.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 4 of 63
4
INTRODUCTION
Most selenoproteins characterized to date are oxidoreductases that utilize thiol/disulfide
exchange reactions to catalyze specific reactions.1 In addition, many selenoproteins have been
shown to have peroxidase activity. This peroxidase activity is due to the “Janusꢀfaced” nature of
selenium;2 the highly nucleophilic selenolate of selenocysteine (Sec)d is readily oxidized by
hydrogen peroxide (H2O2) to form a Secꢀselenenic acid, which is strongly electrophilic. Secꢀ
selenenic acid is then either rapidly reduced by resolving enzymic cysteine (Cys) residues, or by
exogenous thiols like glutathione (GSH).1
Dimedone (5,5ꢀdimethylcyclohexaneꢀ1,3ꢀdione), a small molecule 1,3ꢀdiketone, is
frequently used to detect protein sulfenic acids, as dimedone will alkylate Cysꢀsulfenic acids
forming a stable adduct that can be detected by mass spectrometry.3 It has also found use in
labeling Secꢀselenenic acids of selenoproteins.4ꢀ5 However, in the case of selenoprotein S, the
selenenic acid could only be alkylated when a resolving Cys residue was mutated to serine (Ser),
otherwise rapid formation of a selenosulfide bond prevented alkylation of the selenenic acid with
dimedone.4
Mammalian thioredoxin reductase (mTrxR) is a selenoenzyme whose primary catalytic
function is to reduce the oxidized form of thioredoxin (Trx), a small dithiolꢀcontaining protein
that serves to maintain cellular redox homeostasis.6 The enzyme also has weak peroxidase
activity due to the presence of Sec as discussed above.7 Presumably, this peroxidase activity is
due to the reaction of the Secꢀselenolate with H2O2 to form a Secꢀselenenic acid, which can then
be resolved by an adjacent Cys residue to form a selenosulfide bond as shown in Figure 1(A). If
the lifetime of the Secꢀselenenic acid is long enough it could theoretically be alkylated by
dimedone [Fig. 1(B)], causing inhibition of the enzyme. Similar to the study done with
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 5 of 63
Protein Science
5
selenoprotein S, we also were unable to use dimedone to alkylate the selenenic acid of wild type
mTrxR as shown by the absence of inhibition of the enzyme in the presence of dimedone.4,8
As in the case of selenoprotein S, the lack of alkylation by dimedone could be due to
rapid resolution of the Secꢀselenenic acid by the adjacent, resolving Cys residue [Fig. 2(A)], but
other possibilities exist as well. These other possibilities are: (i) the Secꢀselenenic acid is
alkylated by dimedone, but the alkylation is rapidly reversed by the resolving Cys residue [Fig.
–
2(B)], and (ii) the Sec residue is over oxidized to a Secꢀseleninic acid (SecꢀSeO2 ) [Fig. 2(C)]. If
the reactivity of Secꢀseleninic acid is identical to Cysꢀsulfinic acid, then it is expected that
dimedone would be unreactive towards Secꢀseleninic acid and alkylation would not occur.
However, the reactivity of seleninic acid with dimedone is heretofore unknown
In this study, we explored these three options by studying the reaction of dimedone with
model peptides that correspond to the Cꢀterminal tail of mTrxR, which contains the Sec residue.
We found that the peptide HꢀPTVTGCUGꢀOH, containing the adjacent, resolving Cys residue,
could not be alkylated with dimedone. We were, on the other hand, able to alkylate a “mutant”
peptide (HꢀPTVTGAUGꢀOH) with dimedone in which the resolving Cys residue was replaced
with alanine (Ala). However, this alkylation could be reversed by addition of reducing agents
such as βꢀmercaptoethanol (βME) and tris(2ꢀcarboxyethyl)phosphine (TCEP). Surprisingly,
dimedone alkylation could also be reversed by adding disulfides such as hydroxyethyldisulfide
(HED) and cystine to the alkylated peptide. The diselenideꢀcontaining compound selenocystine
also reversed dimedone alkylation.
An important finding of this study is that dimedone reacts with Secꢀseleninic acid, unlike
the corresponding reaction of dimedone and Cysꢀsulfinic acid. Reaction of dimedone with model
seleninic acids resulted in formation of diselenides and a dimedone dimer as a result of a selenoꢀ
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 6 of 63
6
Pummerer reaction.10ꢀ13 We were able to detect this dimeric form of dimedone when higher
concentrations of H2O2 were used to oxidize the mutant peptide, HꢀPTVTGAUGꢀOH, in the
presence of dimedone. This result may lead to the development of a method to detect Secꢀ
seleninic acid, a highly transient species, in selenoproteins.
MATERIALS AND METHODS
General methods
Unless otherwise stated, all nonꢀaqueous reactions were carried out in either ovenꢀdried
or flameꢀdried glassware. All commercially available starting materials were purchased from
Aldrich, Fischer Scientific, or Acros Organics, and used as received. Analytical thinꢀlayer
chromatography (TLC) was performed on Millipore TLC Silica gel 60 F254 silica gel plates with
UV indicator. Visualization was accomplished by irradiation under a 254 nm UV lamp or stained
with either an aqueous solution of ceric ammonium molybdate, potassium permanganate, or
phosphomolybdic acid. Flash chromatography was performed using a forced flow of the
indicated solvent system on silica gel (32ꢀ63 ꢀm particle size). Removal of solvents was
accomplished on a Büchi Rꢀ114 rotary evaporator and compounds were further dried under a
high vacuum line. Elemental analysis was performed by Atlantic Microlab, Inc (Norcross, GA).
All 1H NMR data was conducted at ambient temperatures and recorded on a Varian Unity
Inova (500 MHz) or Bruker ARX (500 MHz) spectrometer. 13C and 77Se NMR spectra were
recorded on a Bruker ARX spectrometer (125 MHz and 95 MHz, respectively). Chemical shifts
for 1H NMR, 13C NMR, and 13C{1H} NMR, are reported in parts per million (ppm) at 25°C
relative to chloroform or methanol (δ = 7.26 ppm or 3.31 ppm for 1H NMR; δ = 77.16 ppm or
49.00 ppm for 13C NMR, respectively). Data are reported as follows: chemical shift, multiplicity
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 7 of 63
Protein Science
7
(s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, br = broad, m = multiplet), coupling
constants (Hz), and number of protons. Direct infusion mass spectrometry was executed using an
ABI Sciex 4000QTrap Pro LCMS in positiveꢀESI or APCI mode. Liquid chromatographyꢀmass
spectrometry (LCMS) was executed using an ABI Sciex 4000QTrap Pro LCMS equipped with a
C18 column in positiveꢀESI mode.
All analytical HPLC chromatography was carried out on a Shimadzu analytical HPLC
system which utilized LCꢀ10AD pumps, a SPDꢀ10A UVꢀVis detector, and a SCLꢀ10A
controller. A Shimadzu Shimꢀpack™ VPꢀODS C18 analytical column (4.6 ꢁm pore size, 150 ×
4.6 mm) was used in these separations. Beginning with 100% Buffer A, a 1.4 mL/min gradient
elution increase of 1% Buffer B/min for 50 min was used for all peptide chromatograms (Buffer
A = 0.1% TFA/H2O; Buffer B = 0.1% TFA/ACN). Peptide elution profiles were detected at 214
nm and 254 nm.
Production of thioredoxin reductase by semisynthesis
The wild type mTrxRꢀGCUGe enzyme and the mutant mTrxRꢀGGUG enzyme were
produced by protein semisynthesis using well described methods published by our laboratory.14
Attempt to inhibit thioredoxin reductase by dimedone treatment in presence of H2O2
Wild type mTrxRꢀGCUG (10 nM) or mutant mTrxRꢀGGUG (50 nM), were incubated
with 200 ꢁM NADPH and 1 mM EDTA in 100 mM potassium phosphate buffer, pH 7.0 for 5
min. After 5 min, 500 ꢁM dimedone or the same volume of methanol (control) was added to the
reaction mixture along with 300 ꢁM H2O2. The reactions were incubated for 25 min followed by
the addition of catalase (170 nM) for 5 min to quench the H2O2. Enzyme activity was then
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 8 of 63
8
initiated by the addition of 90 ꢁM selenocystine, and enzyme activity was determined
spectrophotometrically by following the consumption of NADPH at 340 nm.
Peptide Synthesis
Peptides were synthesized as previously described using 2ꢀchlorotritylchloride resin and
standard Fmocꢀsolid phase peptide synthesis protocol.14 Once synthesized, peptides were cleaved
from the resin using a cocktail containing 96:2:2 trifluoroacetic acid (TFA), triisopropylsilane
(TIS), H2O, and 2 equivalents of dithionitropyridine (DTNP).15,16 Following cleavage, the
volume was reduced by evaporation under a stream of N2 and peptide was precipitated using
cold Et2O. Once dry, the peptides were suspended in H2O with minimal acetonitrile, frozen (ꢀ
78°C), lyophilized, and purified via preparatory HPLC. Peptide composition was analyzed using
direct infusion APCI mass spectrometry.
Labeling of model Sec-containing peptides with dimedone
Peptides were synthesized with the Sec residue initially protected as the pꢀmethoxybenzyl
(Mob) derivative. The Mob group on Sec was replaced with a 5ꢀnitroꢀ2ꢀpyridinesulfenyl (5ꢀ
Npys) group by addition of DTNP to the cleavage cocktail.15 Upon addition of DTNP, the Mob
protecting group on Sec is replaced with 5ꢀNpys. Attack by the adjacent Cys residue of the wild
type peptide (HꢀPTVTGCUGꢀOH) onto the selenium atom of the Sec(5ꢀNpys) group results in
the rapid formation of an intramolecular selenosulfide bond.
The oxidized, wild type peptide (5 ꢀM) was reduced with 50 ꢀM tris(2ꢀ
carboxyethyl)phosphine (TCEP) in 100 mM potassium phosphate pH 7.0 for 15 min, followed
by incubation with 50 ꢀM or 500 ꢀM dimedone and 5 ꢀM H2O2 at room temperature while being
shaken at low speed for 25 min. The samples were then analyzed via LCMS.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 9 of 63
Protein Science
9
As a control, the same peptide (5 ꢀM) was reduced with 50 ꢀM TCEP in 100 mM
potassium phosphate buffer, pH 7.0 for 15 min. After, 10 mM iodoacetic acid was added and the
reaction was shaken in the dark at room temperature for 25 min. The samples were then analyzed
via LCMS.
In the mutant peptide, HꢀPTVTGAUGꢀOH, the Sec residue was protected as Sec(5ꢀ
Npys), which in this form is less reactive towards oxidation. In order to generate a free selenol in
situ, the 5ꢀNpys protecting group was removed by addition of ascorbate in 100 mM potassium
phosphate pH 7.0 for 30 min at 37°C as we have previously reported.17 The ratio of ascorbate to
peptide (0.9:1) was subꢀstoichiometric to ensure that no ascorbate remained after deprotection
that could potentially scavenge H2O2 or reduce oxidized Sec species during the experiment.
Following deprotection, the peptide (5 ꢀM) was incubated with 50 ꢀM dimedone and 5 ꢀM
H2O2 at room temperature while being shaken at low speed for 25 min. The reactions were then
analyzed using LCMS as described above.
Demonstrating the lability of the Sec-dimedone label
Peptides were labeled with dimedone as described above. Following the 25 min
incubation with 50 ꢀM dimedone and 5 ꢀM H2O2, βꢀmercaptoethanol (βME, 10x), TCEP (10x),
cystine (1000x), or selenocystine (1000x) were added to the reaction mixture. The reactions were
incubated at room temperature for 15 min while being shaken at low speed then analyzed via
LCMS.
77Se NMR time course of reactions between dimedone and seleninic acids
A concentrated solution (~150 mM) of benzeneseleninic acid (PhSeO2H) or
methaneseleninic acid (MeSeO2H) in CD3OD was made and an initial 77Se NMR spectrum was
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 10 of 63
10
taken as the initial time point. After, the solution was diluted at room temperature with a stock
solution of dimedone in CD3OD so that the final concentration of PhSeO2H or MeSeO2H was
100 mM and the concentration of dimedone was 200 mM. The solution was immediately
inserted into the spectrometer and 77Se NMR spectra were taken every 3 h for 6 h (PhSeO2H) or
every 15 min for 30 min (MeSeO2H).
LCMS analysis of reaction between dimedone and PhSeO2H
A 10 mL solution of 100 mM PhSeO2H and 200 mM dimedone in CH3OH was made and
the reaction was allowed to proceed with stirring for 48 h at room temperature. An aliquot of the
reaction mixture was taken and analyzed via LCMS as described above.
Attempt to purify dimedone byproducts of reaction between dimedone and PhSeO2H
Dimedone (280.4 mg, 2 mmol) was dissolved in methanol (10 mL) and benzeneseleninic
acid (189.1 mg, 1 mmol) was added. The reaction was stirred at room temperature for 48 hours.
The solvent was evaporated and the crude mixture was purified via flash chromatography (SiO2;
4:1 ethyl acetate/dichloromethane) to yield 2 (see Fig. 9) as a white, crystalline solid, which was
the only product able to be isolated.
1H NMR (500 MHz, chloroformꢀd) δ 2.28 (s, 4H), 2.22 (s, 4H), 1.63 (s, 6H), 1.06 (s, 12H). 13C
NMR (125 MHz, chloroformꢀd) δ 197.64, 160.58, 119.81, 53.04, 41.54, 32.16, 31.63, 28.19,
26.05. MS (pos ESI) m/z 303.4 [(M + H)+, calculated for C19H26O3: 302.41]. Anal. Calcd. for
C19H26O3: C, 75.46; H, 8.67. Found: C, 74.79; H, 8.67.
Formation and detection of dimeric dimedone species upon reaction of mutant peptide and H2O2.
The 5ꢀNpys protecting group of HꢀPTVTGAU(5ꢀNpys)GꢀOH (1 mM) was removed by
incubation with ascorbate at a ratio of 0.9:1 ascorbate:peptide in 100 mM potassium phosphate
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 11 of 63
Protein Science
11
pH 7.0 for 30 min at 37°C as described above. Following deprotection, the peptide was incubated
with 2 mM dimedone and 5 mM H2O2 for 25 min at room temperature while being shaken at low
speed, after which the samples were analyzed via LCMS.
RESULTS
Attempt to inhibit thioredoxin reductase by dimedone treatment in presence of H2O2
To address the question of whether the adjacent Cys residue of mTrxRꢀGCUG interferes
with alkylation of Secꢀselenenic acid by dimedone, we constructed a mutant enzyme in which
the adjacent Cys residue was replaced by Gly (mTrxRꢀGGUG). Both the wild type enzyme,
mTrxRꢀGCUG, and mutant enzyme, mTrxRꢀGGUG, were subjected to a selenocystine reductase
assay, which measures each enzyme’s ability to reduce selenocystine after exposure to H2O2 and
dimedone.18 In this assay, the % selenocystine reductase activity remaining of each enzyme after
exposure to H2O2 and dimedone is indicative of how much Secꢀselenenic acid is being alkylated
by dimedone. The results from this experiment are shown in Table 1. The data show that the wild
type mTrxRꢀGCUG enzyme retained full selenocystine reductase activity, which is identical to
our previous result.8 However, the mutant mTrxRꢀGGUG enzyme retained only 74% of
selenocystine reductase activity after being exposed to H2O2 and dimedone. We interpret this
result as the Secꢀselenenic acid of the mutant enzyme mTrxRꢀGGUG being alkylated with
dimedone, causing inhibition of selenocystine reductase activity. In this case, the Secꢀselenenic
acid of the mutant enzyme could be alkylated because there was no resolution of Secꢀselenenic
acid by the adjacent Cys residue.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 12 of 63
12
We believe the reason for the lack of complete inhibition of the mutant enzyme by
dimedone is due to the presence of the Nꢀterminal redox center of the enzyme, which contains
two redox active Cys residues that may be able to reverse the alkylation.
Labeling of model Sec-containing peptides with dimedone
The peptides that we chose to synthesize were based on the Cꢀterminal amino acid
sequence of mTrxR. We either synthesized the wild type peptide, HꢀPTVTGCUGꢀOH, or mutant
peptide, HꢀPTVTGAUGꢀOH. In the case of the mutant peptide, the antepenultimate Cys residue
has been mutated to Ala to avoid resolution of Secꢀselenenic acid, as would occur if the adjacent
Cys was present in the peptide, resulting in the formation of a selenosulfide bond [Fig. 1(A)].
First, we performed a control experiment with the oxidized, wild type peptide. We
reduced the wild type peptide with a 10ꢀfold excess of TCEP and subsequently alkylated it with
iodoacetic acid. The results showed that both the Cys and Sec residue of the peptide could be
carboxymethylated (Fig. S1). However in a similar experiment, we were unable to alkylate the
wild type peptide with either 50 ꢀM dimedone (10ꢀfold excess) or 500 ꢀM dimedone (100ꢀfold
excess) in the presence of H2O2, identical to our results with the fullꢀlength enzyme (Fig. S2, S3).
The fact that we were able to dialkylate the peptide with iodoacetic acid shows that we were able
to effectively reduce the selenosulfide bond with TCEP, making both the sulfur and selenium
susceptible to oxidation by H2O2 and subsequent labeling by dimedone. The lack of dimedone
alkylation shows that it must be the specific Se–C bond of the Secꢀdimedone adduct that is labile
for the reasons we discuss in Figure 2, as the carboxymethylation of Sec by iodoacetic acid was
not reversed by the resolving Cys residue.
A difficulty in our experimental design with the mutant peptide that had to be overcome
was the generation of free Secꢀselenol. Alkyl selenols do not exist in solution for a long duration
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 13 of 63
Protein Science
13
13due to their tendency to oxidize to the diselenide form.19 To mitigate this problem, we utilized
a strategic protecting group, 5ꢀNpys, on Sec, that allowed us to rapidly generate Secꢀselenol in
situ, which could be further oxidized to Secꢀselenenic acid by H2O2 and alkylated with dimedone
as shown in Figure 3. The selenosulfide bond between the selenium of Sec and the sulfur of the
5ꢀNpys group is readily cleaved by ascorbate, generating dehydroascorbic acid in the process.17
Therefore, before oxidizing the Sec peptide in the presence of dimedone, the relevant selenol
form of Sec was generated by incubating the 5ꢀNpysꢀprotected Secꢀpeptide with ascorbate. This
approach does not depend on the use of an exogenous thiol to generate the selenol. The presence
of exogenous thiol would have greatly complicated the interpretation of the results, as the added
thiol could potentially reverse alkylation of Sec by dimedone.
The results of our attempts to alkylate the Secꢀselenenic acid of peptide HꢀPTVTGAUGꢀ
OH with dimedone are displayed in Figure 4. The extracted ion chromatogram in Figure 4(A)
shows that the vast majority of peptide HꢀPTVTGAUGꢀOH is oxidized to the Secꢀselenenic acid
and subsequently alkylated with dimedone. Peptide that was not labeled was oxidized to the
diselenide, which is the product of the disproportionation of two selenenic acids.19 Thus, the 10ꢀ
fold excess of dimedone relative to peptide used in the experiment was sufficient for the
alkylation reaction to outcompete disproportionation. We note that, in both cases, the mass
spectra shown in Figure 4(B) and 4(C) clearly display the characteristic isotope pattern of
seleniumꢀcontaining peptides, which dispels any doubt that the signal corresponding to the
dimedoneꢀalkylated HꢀPTVTGAUGꢀOH is artifactual.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 14 of 63
14
Demonstrating the lability of the Sec-dimedone label
After successfully alkylating the Secꢀselenenic acid of peptide HꢀPTVTGAUGꢀOH with
dimedone, we desired to address the question raised in Figure 2(B), that is, could the reason why
the wild type form of mTrxR retains peroxidase activity even in the presence of dimedone be due
to the fact that the dimedone label is being removed by the adjacent, resolving Cys residue? To
address this question, we subjected mutant peptide HꢀPTVTGAUGꢀOH to the same conditions
used to alkylate the Secꢀselenenic acid with dimedone, then immediately incubated the alkylated
peptide with 10ꢀfold molar excesses of βME or TCEP for 15 min. Figure 5 shows the results
from these experiments, which demonstrate that under all reducing conditions employed, the
Secꢀdimedone label is labile. In the case of 10x βME, more than half of the dimedone label was
removed and the peptide was detected as the βME adduct (Fig. 5). When 10x TCEP was used as
the reducing agent, 50ꢀ60% of the label was also removed. The peptide was detected in the
reduced form (Fig. 5), but the TCEP adduct was not detected. When 1000x βME was used, no
dimedoneꢀlabeled peptide HꢀPTVTGAUGꢀOH was detected and only the βME adduct was
observed (data not shown). For comparison, experiments conducted by Carroll et al. showed that
the Cysꢀdimedone label of a model dipeptide was stable over 12 h to similar reducing conditions
of 10ꢀfold molar excesses of dithiothreitol (DTT), glutathione (GSH), or TCEP.20
To further validate our experiment showing that addition of βME results in removal of
the dimedone label, we did a control experiment in which hydroxyethyl disulfide was added
instead. To our surprise, addition of this disulfide also resulted in removal of the dimedone label
(data not shown). We then repeated this experiment using the more relevant disulfide compound
cystine instead, and this disulfide also resulted in removal of the dimedone label and formation
of a cysteinylꢀpeptide adduct. We reasoned that addition of a diselenide should enhance removal
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 15 of 63
Protein Science
15
of the dimedone label, so selenocystine was added as a removal agent. Indeed, we found that
addition of the diselenide enhanced the amount of dimedone label removed from the peptide (see
the bottom two panels of Fig. 5).
Reaction of dimedone with model seleninic acids
The reactivity of dimedone towards seleninic acids has not been reported in the literature
so far as we are aware. While we showed that one possible reason mTrxR is not inhibited by
dimedone in the presence of H2O2 is due to the removal of the Secꢀdimedone adduct by thiols
(Fig. 2(B)], the question raised in Figure 2(C) still remained unaddressed, could dimedone be
reacting with a Secꢀseleninic acid? We hypothesized that dimedone would react with seleninic
acids because of their high Seꢀelectrophilicity. Seleninic acids, unlike sulfinic acids, are highly
electrophilic and can be rapidly reduced by thiols and other reducing agents.2
The reactivity of dimedone towards seleninic acids was first explored in a qualitative,
model experiment, where a solution of 100 mM PhSeO2H and 200 mM dimedone was prepared
in methanol. After approximately 15 min, the solution turned from clear, colorless to an intense
yellow color, indicating a reaction between the two compounds. The observation of this yellow
color led us to hypothesize that one product of the reaction could be diphenyl diselenide, which
has a yellow color in solution.
In order to further test this hypothesis, we employed a 77Se NMR time course experiment
where the reaction between PhSeO2H and dimedone in CD3OD was followed for 6 h in 3 h
intervals. The results from this experiment are shown in Figure 6. As shown in the figure, a
resonance at δ = 463 ppm appeared within 3 h of reaction time and remained the only observable
resonance after 6 h. A control 77Se NMR spectrum of the pure compound confirmed that this
resonance corresponds to diphenyl diselenide, which further supports our hypothesis that
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 16 of 63
16
dimedone is indeed reducing PhSeO2H (+2 oxidation state) to the diselenide form (ꢀ1 oxidation
state). Figure S4 shows the 77Se NMR spectrum of diphenyl diselenide.
Next, in order to show that dimedone would react not only with aromatic seleninic acids,
but also with alkyl seleninic acids such as Sec, we monitored the reaction of commercially
available MeSeO2H with dimedone in CD3OD using the same 77Se NMR time course
experiment.
As shown in Figure 7, dimedone quantitatively reduces MeSeO2H to dimethyl diselenide
(δ = 263 ppm) after only 30 min of reaction time (see Fig. S5 for the control 77Se NMR spectrum
of dimethyl diselenide).
As a control, the reactivity of dimedone towards benzenesulfinic acid (PhSO2H), the
sulfurꢀcontaining analog of PhSeO2H, was explored in a 1H NMR time course experiment (see
Fig. S6). As expected, there was no observable reaction between dimedone and PhSO2H even
after 7 days of reaction time. Furthermore, since PhSeO2H is a weak acid (pKa is 4.79), we
conducted a control experiment designed to show that dimedone was, in fact, undergoing a redox
reaction with the model seleninic acids, and not just an acidꢀbase reaction.21 The reactivity of
dimedone in acidic media generated by pꢀtoluenesulfonic acid (pꢀTsOH) was determined in a
similar 1H NMR time course experiment and after 24 h there was no observable change in the 1H
NMR spectrum (Fig. S7), providing further evidence that dimedone is, indeed, participating in a
redox reaction with the model seleninic acids.
While the identities of the seleniumꢀcontaining products of the reaction between
dimedone and either PhSeO2H or MeSeO2H were readily revealed by our 77Se NMR time course
experiments, the fate of dimedone in the redox reaction could not be determined by 77Se NMR.
In an attempt to determine the oxidation products of dimedone in its reaction with seleninic
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 17 of 63
Protein Science
17
acids, we conducted 1H NMR time course experiments monitoring the reaction between either
PhSeO2H or MeSeO2H, and dimedone in CD3OD (see Fig. S8, S9). However, these experiments
did not aid in product identification because only new alkyl 1H NMR resonances, slightly
downfield of the original resonances, were observed during the time course, and these
resonances were not informative.
Despite this result, we hypothesized that the seleninic acids were undergoing a variation
of the Pummerer reaction, as shown in Figure 8.10ꢀ13 This hypothesis was based on the fact that
the seleninic acid was being reduced to the diselenide, which means that dimedone must be
oxidized during the reaction. In the first step of the proposed mechanism, the nucleophilic carbon
atom of dimedone attacks the electrophilic selenium atom of the seleninic acid with subsequent
elimination of a molecule of water, resulting in the formation of a selenoxide intermediate. After
eliminating another molecule of water, the electrophilic selenonium species is attacked by
another molecule of dimedone to produce a selenide. Last, elimination of selenolate anion (pKa
~ 5.5)22 results in the formation of dimeric dimedone species 1a. The selenolate anion formed is
readily oxidized to the diselenide by dissolved O2 in the buffer. The generation of 1a would
coincide with the observation that new downfield alkyl resonances appeared during our 1H NMR
time course experiments.
In order to verify our hypothesis that 1a is a product of our proposed selenoꢀPummerer
reaction, we subjected an aliquot of the reaction between dimedone and PhSeO2H for high
performance liquid chromatographyꢀmass spectrometry (LCMS) analysis. Figure 9(A) displays
the HPLC total total ion current (TIC) of the reaction mixture, which shows the elution of two
analytes with retention times of 18.3 min and 26.7 min, respectively. The resulting mass spectra
obtained from the eluted analyte at t = 18.3 min is shown in Figure 9(B). The spectrum shows a
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 18 of 63
18
peak at m/z = 277.2, which was identified as the [M+H]+ ion of 1a. In addition, the observation
of a more intense peak at m/z = 309.3, an increase of 32 Da, supports the existence of 1a as this
mass corresponds to the methanol adduct (1b). The alkene center of 1a is highly electrophilic
and is unstable in nucleophilic solvents. As a result, 1a could not be isolated through
employment of standard chromatographic conditions on SiO2, a result likely attributable to the
highly electrophilic character of this alkene.
The mass spectrum obtained from the analyte observed at t = 26.7 min is shown in Figure
9(C). It shows only one main ion at m/z = 303.4, which we attribute to the [M+H]+ ion of the
condensation product between two dimedone molecules and acetone (2).23 Acetone could be
present due to either breakdown of dimedone during the reaction or contamination of the reaction
glassware. Purification of 2 on SiO2 yielded 1H, 13C{1H}, and 13CꢀDEPT135 NMR spectra, as
well as elemental analysis data that matched the compound (see Fig. S10ꢀS14 for
characterization data of 2). One can envision a mechanism to produce 2 without direct redox
interaction with the selenium atom of a seleninic acid. As 2 is the minor product of the reaction
[see Fig. 9(A)], it was of little significance to this study.
The presence of 1a as a chemical signature of seleninic acid in a model peptide
Recently a molecular probe was developed to detect protein Cysꢀsulfinic acids, the
sulfurꢀcontaining analog of Secꢀseleninic acids.24 However, no similar chemical probe is known
to detect the presence of Secꢀseleninic acid. Our findings detailed above led us to pursue the use
of 1a as a chemical footprint for the detection of Secꢀseleninic acid. We envisioned an assay in
which a Secꢀcontaining peptide or protein would be exposed to H2O2 in the presence of
dimedone. This could potentially generate a Secꢀseleninic acid in situ that could react with
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 19 of 63
Protein Science
19
dimedone to produce 1a, which could then be detected by a LCMS experiment similar to the one
presented in Figure 9.
To this end, we conducted an experiment where our model Secꢀpeptide, HꢀPTVTGAUGꢀ
OH, was exposed to a 5 molar excess of H2O2 while incubated with a 2 molar excess of
dimedone. The result from this experiment is shown in the top panel of Figure 10. The extracted
ion chromatogram shows that, under these conditions, 1a was produced from dimedone’s
reaction with the Secꢀseleninic acid of peptide HꢀPTVTGAUGꢀOH. Similar to the case where
methanol was the solvent, H2O adds to the electrophilic alkene center of 1a to form adduct 1c
when water is the solvent (see Fig. 10). In order to demonstrate that the presence of peptide Hꢀ
PTVTGAUGꢀOH is required for the generation of 1a, we conducted a control experiment where
dimedone was reacted with H2O2 alone. A small amount of 1a is present in this control
experiment as shown in the bottom panel of Figure 10. However, during the course of this
experiment, an unknown, contaminating seleniumꢀcontaining compound, eluted from the column
(identified by the characteristic isotope pattern of selenium). This contaminant is most likely
present from previous exposure of the LCMS system to selenopeptides. We hypothesized that the
excess H2O2 injected into the LCMS reacted with this seleniumꢀcontaining compound to
generate a seleninic acid in situ during separation on the LC column, thereby generating 1a upon
its reaction with dimedone. Further control experiments using 1H NMR and directꢀinfusion ESI
mass spectrometry provided evidence for this claim as no reaction between dimedone and H2O2
was observed during these experiments (see Fig. S15 for 1H NMR control; see Fig. S16 for
directꢀinfusion ESI mass spectrometry control). Therefore, the selectivity of our proposed assay
for seleninic acids is confirmed in that the generation of 1a occurs in neither the presence of
H2O2 nor sulfinic acids.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 20 of 63
20
DISCUSSION
As described in Figure 2, here we have revisited the quandary of why mTrxR cannot be
inhibited by dimedone, even though a Secꢀselenenic acid, a strong electrophile, must be formed
upon reaction of the enzyme with H2O2.8 Our curiosity led us to delve into this problem based on
two key observations that are described in the following two sections.
Dimedone and Sec-selenenic acids
The first key observation came in the form of a literature search for the reported pKa
value of the α-hydrogens of dimedone, which we found to be ~5.9 This low pKa renders
dimedone a suitable leaving group in SN2ꢀlike reactions. Further motivation was found in the
form of multiple reports detailing the lability of both α-keto sulfides and selenides towards thiols
and selenols.25ꢀ31 The Secꢀdimedone label comprises an α-diketo selenide, which we
hypothesized would be even more labile than the reported α-monoketo selenides. Coupled with
the fact that CꢀSe bonds are weaker than CꢀS bonds due to their lowerꢀlying σ* orbital,32 we
thought that the Secꢀdimedone bond might be labile to nucleophiles. This idea was supported by
our initial experiment where we showed that the wild type mTrxRꢀGCUG enzyme retained full
selenocystine reductase activity in the presence of H2O2 and dimedone, whereas the mutant
enzyme mTrxRꢀGGUG (no resolving Cys) only retained 74% of selenocystine reductase activity
under the same conditions. However, it was still unclear whether the wild type enzyme was
retaining activity due to the mechanism shown in Figure 2(A) (resolution of Secꢀselenenic acid
by resolving Cys) or the mechanism in Figure 2(B) (removal of dimedone label by resolving
Cys).
To provide thorough evidence that the Secꢀdimedone adduct is labile and that the
mechanism shown in Figure 2(B) is plausible, we turned to a peptide model to facilitate our
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 21 of 63
Protein Science
21
analysis. We explicitly chose to remove the resolving Cys residue from our model wild type,
peptide HꢀPTVTGCUGꢀOH. The reason behind this was to eliminate the possibility of
selenosulfide bond formation through the mechanism shown in Figure 2(A). Reducing agents
were added after the peptide was alkylated with dimedone to mimic the resolving Cys in the
enzyme and the results show that, indeed, the Secꢀdimedone bond is labile.
Possible mechanisms explaining the removal of the dimedone label on HꢀPTVTGAUGꢀ
OH are shown in Figure 11. Whether βME [Fig. 11(A)], TCEP [Fig. 11(B)] or
selenocystine/cystine [Fig. 11(C)] are used as removal reagents, the selenium atom of the Secꢀ
dimedone adduct acts as the electrophile in mechanisms similar to the one shown in Figure 2(B),
except here exogenous nucleophiles are used instead of the resolving Cys residue of the
peptide/enzyme. The 10ꢀfold excess of βME and TCEP used in this experiment is conservative
when considering the theoretical effective molarity of the adjacent, resolving Cys residue in the
peptide/enzyme.33ꢀ35 The surprising result that cystine (a disulfide compound) and selenocystine
(a diselenide compound) were able to partially remove the Secꢀdimedone label [Fig. 11(C)]
demonstrates the lability of the Secꢀdimedone adduct.
An alternative mechanism to that shown in Figure 11(C) can also be envisioned in which
the selenide of the Secꢀdimedone acts as nucleophile to attack an electrophilic disulfide (cystine)
or diselenide (selenocystine). The use of a selenide as a nucleophile is not unprecedented.36
Indeed, it is this characteristic of selenides that led to the development of a protecting group
strategy that allows for the facile deprotection of Sec residues protected with the pꢀ
methoxybenzyl group by our lab.16
While our data provide evidence that the mechanism shown in Figure 2(B) may explain
why the wild type mTrxRꢀGCUG enzyme is not inhibited by dimedone in the presence of H2O2,
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 22 of 63
22
the data also demonstrate that dimedone may not be the most suitable molecular probe to identify
Secꢀselenenic acids in proteins. In a previous report, Carroll et al. showed that the dimedone
label of a Cysꢀcontaining dipeptide was stable over 12 h when exposed to reducing conditions of
10x DTT, TCEP, or GSH in potassium phosphate buffer, pH 7.4.20 In a subsequent report, the
authors studied the stability of linear, nucleophilic probes (designed to alkylate Cysꢀsulfenic
acid) to similar reducing conditions.37 Interestingly, they found that, for the most part, there was
no direct correlation between leaving group potential (pKa) and stability of the Cysꢀnucleophile
label. The weaker C–Se bond relative to the C–S bond most likely renders the Secꢀdimedone
adduct inherently labile when compared to the Cysꢀdimedone adduct.
Therefore, caution must be taken when attempting to use dimedone to alkylate Secꢀ
selenenic acid. Any soft nucleophile, such as a thiol or phosphine, that is present during or after
alkylation conditions has the ability to remove the Secꢀdimedone label as shown in Figure 11.
Intracellular reducing agents such as GSH, or denaturing agents such as DTT, may cause
removal of the Secꢀdimedone label during in vivo and in vitro assays, respectively, potentially
leading to false negatives. Last, special attention must be paid to enzymes that possess free Cys
residues near the site of Secꢀdimedone alkylation, as in the case with mTrxR and selenoprotein
S. These Cys residues will interfere with Secꢀselenenic acid alkylation by either (i) reducing the
Secꢀselenenic acid, resulting in a selenosulfide bond [Fig. 2(A)], or (ii) removing the dimedone
label [Fig. 2(B)].
Dimedone and Sec-seleninic acids
Our second key observation, which led us to explore dimedone’s reactivity towards
seleninic acids, was the realization of the explicitly opposite nature of seleninic acids when
compared to sulfinic acids. Sulfinic acids are widely known as reducing agents, whereas
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 23 of 63
Protein Science
23
seleninic acids are considered weak oxidizing agents.2,38ꢀ40 The reducing (nucleophilic) property
of sulfinic acids is exemplified when considering the nature of the molecular probe that was
recently developed for Cysꢀsulfinic acids (NOꢀBio): the probe is electrophilic and is the recipient
of electrons from a Cysꢀsulfinic acid.24 On the other hand, the highly oxidative (electrophilic)
nature of seleninic acids allows them to be readily reduced by thiols and other, mild reducing
agents such as ascorbate.2,41ꢀ44
Both PhSeO2H and MeSeO2H are inexpensive and commercially available and served us
well as model compounds to investigate dimedone’s reactivity towards seleninic acids. The
results from our 77Se NMR time course experiments and our LCMS experiments using PhSeO2H
and MeSeO2H showed that these seleninic acids were being reduced to the corresponding
diselenide compounds and, as a result, were oxidizing dimedone to the dimeric form 1a in the
selenoꢀPummerer reaction detailed in Figure 8.
We were initially disappointed that dimedone did not “tag” the model seleninic acids,
forming a stable adduct that could be detected and used to identify a protein Secꢀseleninic acid.
However, the formation of 1a led us to pursue the development of a potential assay for Secꢀ
seleninic acids. In the assay, 1a acts as a detectable chemical footprint indicating that a transient
Secꢀseleninic acid had formed when Sec was oxidized with H2O2. We had high confidence
moving forward with the development with the knowledge that both sulfinic acids and H2O2
itself are unreactive to dimedone, giving the potential assay high selectivity for Secꢀseleninic
acid.
The validity of the assay was confirmed during our experiments with model peptide Hꢀ
PTVTGAUGꢀOH. In this case, the Secꢀseleninic acids were generated in situ to mimic
conditions that would occur in both in vitro and in vivo assays. We were able to detect the
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 24 of 63
24
formation of 1a via LCMS after exposing the peptide HꢀPTVTGAUGꢀOH to a 5ꢀmolar excess of
H2O2 and a 2ꢀmolar excess of dimedone. A large excess of dimedone relative to peptide is not
required for the formation of 1a, indicating reasonably fast reaction kinetics for assay viability.
This report marks the first time a molecular probe has been used to identify the presence of a
Secꢀseleninic acid.
To date, the transient seleninic acid species has been detected only by 77Se NMR,44ꢀ50 Xꢀ
ray crystallography,49,51ꢀ53 and mass spectrometry.54ꢀ55 However, these methods of detection are
not ideal. For instance, 77Se NMR is a very insensitive technique with sensitivity similar to
13C{1H} NMR, owing to its slightly higher natural abundance but lower gyromagnetic ratio
relative to 13C.56 Therefore, detection of protein Secꢀseleninic acids via this method is difficult.
Moreover, Xꢀray crystallography is only amenable to those few Secꢀseleninic acids that are
stabilized by the protein microenvironment, as in the case with selenosubtilisin and glutathione
peroxidase.52ꢀ53 Finally, detection of protein Secꢀseleninic acids by mass spectrometry is
unreliable due to their high reactivity. The shortcomings of these methods to detect Secꢀseleninic
acids render the potential assay we have developed, using 1a as a detectable chemical footprint,
advantageous for Secꢀseleninic acid detection. Further development is currently underway in our
lab to refine this assay to make it viable for in vivo detection of Secꢀseleninic acids.
CONCLUSIONS
This report has clearly demonstrated the differences in the chemical properties of Cysꢀ
sulfenic acid and Secꢀselenenic acid towards alkylation by dimedone. A Cysꢀdimedone adduct is
much more stable towards nucleophiles compared to a Secꢀdimedone adduct, which is labile
towards nucleophiles. The lability of the Secꢀdimedone adduct is problematic for the detection of
Secꢀselenenic acids under in vivo conditions due to the presence of glutathione. Further
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 25 of 63
Protein Science
25
differences between Sꢀoxides and Seꢀoxides is demonstrated by the finding that Cysꢀsulfinic acid
is unreactive towards dimedone, while Secꢀseleninic acid reacts with dimedone to yield and a
diselenide and a dimedone dimer, the latter forms as a result of the selenoꢀPummerer reaction.
The presence of the dimedone dimer may possibly be used in the future for the detection of
transient Secꢀseleninic acids in enzymes/proteins.
ACKNOWLEDGEMENT
We would like to acknowledge Bruce O’Rourke of the UVM Department of Chemistry for
acquiring the LCMS data.
SUPPORTING INFORMATION AVAILABLE
There are 16 figures in the Supporting Information. Figures S1 shows the extracted ion
chromatogram and the corresponding mass spectrogram of the reaction of the wild type peptide
with TCEP and iodoacetic acid. Figures S2 and S3 shows the extracted ion chromatogram and
the corresponding mass spectrogram of the reaction of the wild type peptide with TCEP with 5
ꢀM H2O2 and 50 ꢀM or 500 ꢀM dimedone, respectively. Figures S4 and S5 show the 77Se NMR
spectra of diphenyl diselenide and dimethyl diselenide, respectively. Figure S6 shows the 1H
NMR time course of the reaction of benzene sulfinic acid with dimedone. Figure S7 shows the
1H NMR spectrum of pꢀTsOH, dimedone, and the solution of both compounds after 1 h. Figures
S8 and S9 show the 1H NMR time courses of the reactions between either benzene seleninic acid
(Figure S8) or methane seleninic acid (Figure S9) with dimedone. Figures S10ꢀS14 display
characterization spectra of compound 2 (1H NMR, 13C NMR, 13CꢀDEPT135 NMR, posꢀESI mass
spectrum, and elemental analysis). Figures S15 and S16 show the control reactions of dimedone
with H2O2 monitored by 1H NMR and LCꢀMS, respectively.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 26 of 63
26
REFERENCES
1. Hondal RJ, Marino SM, Gladyshev VN (2012) Selenocysteine in thiol/disulfideꢀlike
exchange reactions. Antiox Redox Signal 18:1675ꢀ1689.
2. Reich HJ, Hondal RJ (2016) Why nature chose selenium. ACS Chem Biol 11:821ꢀ841.
3. Klomsiri C, Nelson KJ, Bechtold E, Soito L, Johnson LC, Lowther WT, Ryu SꢀE, King SB,
Furdui CM, Poole LB (2010) Use of dimedoneꢀbased chemical probes for sulfenic acid
detection. Methods Enzymol 473:77ꢀ94.
4. Liu J, Rozovsky S (2013) Contribution of selenocysteine to the peroxidase activity of
selenoprotein S. Biochemistry 52:5514ꢀ5516.
5. Liu J, Zhang Z, Rozovsky S (2014) Selenoprotein K form an intermolecular diselenide bond
with unusually high redox potential. FEBS Lett 588:3311ꢀ3321.
6. Lu J, Holmgren A (2014) The thioredoxin antioxidant system. Free Radic Biol Med 66:75ꢀ
87.
7. Zhong L, Holmgren A (2000) Essential role of selenium in the catalytic activities of
mammalian thioredoxin reductase revealed by characterization of recombinant enzymes with
selenocysteine mutations. J Biol Chem 275:18121ꢀ18128.
8. Snider G, Grout L, Ruggles EL, Hondal RJ (2010) Methaneseleninic acid is a substrate for
truncated mammalian thioredoxin reductase: Implications for the catalytic mechanism and
redox signaling. Biochemistry 49:10329ꢀ10338.
9. Wong FM, Keeffe JR, Wu W (2002) The strength of a lowꢀbarrier hydrogen bond in water.
Tetrahedron Lett 43:3561ꢀ3564.
10. Pummerer R (1909) Über Phenylꢀsulfoxyessigsäure. Berichte der deutschen chemischen
Gesellschaft 42:2282ꢀ2291.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 27 of 63
Protein Science
27
11. Pummerer R (1910) Über Phenylsulfoxyꢀessigsäure (II). Berichte der deutschen chemischen
Gesellschaft 43:1401ꢀ1412.
12. Hagiwara H, Kafuku K, Sakai H, Kirita M, Hoshi T, Suzuki T, Ando M (2000) Domino
Michaelꢀseleno Pummerer type reaction (additive seleno Pummerer reaction). J Chem Soc
Perkin Trans 1:2577ꢀ2578.
13. Smith LHS, Coote SC, Sneddon HF, Procter DJ (2010) Beyond the Pummerer reaction:
Recent developments in thionium ion chemistry. Angew Chem Intl Ed 49:5832ꢀ5844.
14. Eckenroth B, Harris K, Turanov AA, Gladyshev VN, Raines RT, Hondal RJ (2006)
Semisynthesis and characterization of mammalian thioredoxin reductase. Biochemistry
45:5158ꢀ5170.
15. Harris KM, Flemer S, Hondal RJ (2007) Studies on deprotection of cysteine and
selenocysteine sideꢀchain protecting groups. J Peptide Sci 13:81ꢀ93.
16. Flemer Jr S, Lacey BM, Hondal RJ (2008) Synthesis of peptide substrates for mammalian
thioredoxin reductase. J Peptide Sci 14:637ꢀ647.
17. SteMarie EJ, Ruggles EL, Hondal RJ (2016) Removal of the 5ꢀnitroꢀ2ꢀpyridineꢀdulfenyl
protecting group from selenocysteine and cysteine by ascorbolysis. J Peptide Sci 22:571ꢀ576.
18. Cunniff B, Snider GW, Fredette N, Hondal RJ, Heintz NH (2013) A direct and continuous
assay for the determination of thioredoxin reductase activity in cell lysates. Analyt Biochem
443:34ꢀ40.
19. Besse D, Siedler F, Diercks T, Kessler H, Moroder L (1997) The redox potential of
selenocystine in unconstrained cyclic peptides. Angew Chem Intl Ed 36:883ꢀ885.
20. Gupta V, Carroll KS (2016) Profiling the reactivity of cyclic Cꢀnucleophiles towards
electrophilic sulfur in cysteine sulfenic acid. Chem Sci 7:400ꢀ415.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 28 of 63
28
21. McCullough JD, Gould ES (1949) The dissociation constants of some monoꢀsubstituted
benzeneseleninic acids. J Am Chem Soc 71:674ꢀ676.
22. Byun BJ, Kang YK (2011) Conformational preferences and pKa value of selenocysteine
residue. Biopolymers 95:345ꢀ353.
23. Naeimi H, Nazifi ZS (2013) A highly efficient nanoꢀFe3O4 encapsulatedꢀsilica particles
bearing sulfonic acid groups as a solid acid catalyst for synthesis of 1,8ꢀdioxoꢀ
octahydroxanthene derivatives. J Nanoparticle Res 15:2026.
24. Lo Conte M, Carroll KS (2012) Chemoselective ligation of sulfinic acids with arylꢀnitroso
compounds. Angew Chem Intl Ed 124:6608ꢀ6611.
25. Michinori O, Wataru F, Atsuko N (1971) The reaction of αꢀcarbonyl sulfides with bases. I.
The reaction between αꢀcarbonyl sulfides with thiolates. Bull Chem Soc Japan 44:828ꢀ832.
26. Oki M, Funakoshi W (1971) The reaction of αꢀcarbonyl sulfides with bases. II. The effect of
variation in the nucleophiles. Bull Chem Soc Japan 44:832ꢀ835.
27. Reich HJ, Renga JM, Reich IL (1975) Organoselenium chemistry. Conversion of ketones to
enones by selenoxide syn elimination. J Am Chem Soc 97:5434ꢀ5447.
28. Takahashi T, Nagashima H, Tsuji J (1978) Preparation of 1ꢀphenylseleno 2ꢀalkanones from
terminal olefins and their application to organic synthesis. Tetrahedron Lett 19:799ꢀ802.
29. Zima G, Barnum C, Liotta D (1980) Synthetic applications of 2ꢀphenylselenenyl enones.
Selective formation of exocyclic or endocyclic enones from a common intermediate. J Organ
Chem 45:2736ꢀ2737.
30. Makoto S, Ryo T, Isao K (1981) Oxidation of olefins into αꢀphenylseleno carbonyl
compounds. Highly regioselective antiꢀMarkownikoff type oxidation of allylic alcohol
derivatives. Bull Chem Soc Japan 54:3510ꢀ3517.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Page 29 of 63
Protein Science
29
31. Arai H, Kasai M (1993) Facile nucleophilic cleavage of selenide with dimedone. Synthesis of
novel 6ꢀdemethylmitomycins. J Organ Chem 58:4151ꢀ4152.
32. McKillop A (2002) MainꢀGroup Metal Organometallics in Organic Synthesis. Pergamon.
33. Jacobson H, Stockmayer WH (1950) Intramolecular reaction in polycondensations. I. The
theory of linear systems. J Chem Phys 18:1600ꢀ1606.
34. Page MI (1973) The energetics of neighbouring group participation. Chem Soc Rev 2:295ꢀ
323.
35. Kirby AJ (1980) Effective molarities for intramolecular reactions. Advan Phys Organ Chem
17:183ꢀ278.
36. Pearson RG, Sobel HR, Songstad J (1968) Nucleophilic reactivity constants toward methyl
iodide and transꢀdichlorodi (pyridine) platinum (II). J Am Chem Soc 90:319ꢀ326.
37. Gupta V, Carroll KS (2016) Rational design of reversible and irreversible cysteine sulfenic
acidꢀtargeted linear Cꢀnucleophiles. Chem Commun 52:3414ꢀ3417.
38. Barton DHR, Brewster AG, Hui RAHF, Lester DJ, Ley SV, Back TG (1978) Oxidation of
alcohols using benzeneseleninic anhydride. J Chem Soc Chem Commun 21:952ꢀ954.
39. Barton DHR, Finet JꢀP, Thomas M (1988) Comparative oxidation of phenols with
benzeneseleninic anhydride and with benzeneseleninic acid. Tetrahedron 44:6397ꢀ6406.
40. Reich H, Jasperse C (1988) Organoselenium chemistry. Preparation and reactions of 2, 4, 6‐
tri‐tert‐butylbenzeneselenenic acid (VI). ChemInform 19:[PAGE #S].
41. Kice JL, Lee TWS (1978) Oxidationꢀreduction reactions of organoselenium compounds. 1.
Mechanism of the reaction between seleninic acids and thiols. J Am Chem Soc 100:5094ꢀ
5102.
John Wiley & Sons
This article is protected by copyright. All rights reserved.

Protein Science
Page 30 of 63
30
42. Ganther HE, Robert Lawrence J (1997) Chemical transformations of selenium in living
organisms. Improved forms of selenium for cancer prevention. Tetrahedron 53:12299ꢀ12310.
43. Ip C, Thompson HJ, Zhu Z, Ganther HE (2000) In vitro and in vivo studies of
methylseleninic acid: evidence that a monomethylated selenium metabolite is critical for
cancer chemoprevention. Cancer Res 60:2882ꢀ2886.
44. Payne NC, Geissler A, Button A, Sasuclark AR, Schroll AL, Ruggles EL, Gladyshev VN,
Hondal RJ (2017) Comparison of the redox chemistry of sulfurꢀ and seleniumꢀcontaining
analogs of uracil. Free Radic Biol Med 104:249ꢀ261.
45. House KL, Dunlap RB, Odom JD, Wu ZP, Hilvert D (1992) Structural characterization of
selenosubtilisin by seleniumꢀ77 NMR spectroscopy. J Am Chem Soc 114:8573ꢀ8579.
46. Sarma BK, Mugesh G (2008) Antioxidant activity of the antiꢀinflammatory compound
Ebselen: A reversible cyclization pathway via selenenic and seleninic acid intermediates.
Chemistry 14:10603ꢀ10614.
47. Bhabak KP, Vernekar AA, Jakka SR, Roy G, Mugesh G (2011) Mechanistic investigations
on the efficient catalytic decomposition of peroxynitrite by ebselen analogues. Organ Biomol
Chem 9:5193ꢀ5200.
48. Prabhu P, Singh BG, Noguchi M, Phadnis PP, Jain VK, Iwaoka M, Priyadarsini KI (2014)
Stable selones in glutathioneꢀperoxidaseꢀlike catalytic cycle of selenonicotinamide
derivative. Organ Biomol Chem 12:2404ꢀ2412.
49. Singh VP, Poon Jꢀf, Butcher RJ, Engman L (2014) Pyridoxineꢀderived organoselenium
compounds with glutathione peroxidaseꢀlike and chainꢀbreaking antioxidant activity.
Chemistry 20:12563ꢀ12571.
John Wiley & Sons
This article is protected by copyright. All rights reserved.